Sensors and Electron Devices Directorate Research Areas

Signal and Image Processing

Advisor: Nasrabadi, Nasser M. (nasser.m.nasrabadi@us.army.mil 301-394-0806)
Adelphi, MD

Key words: Signal and Image Processing

We explore digital image-processing algorithms to enhance the automated recognition of targets (e.g., military vehicles) in visible, infrared, Hyperspectral or radar imagery. These targets could be at different ranges and aspects, and in various types of clutter. Over the past few years, we have developed several automatic target recognition (ATR) detectors and classifiers for forward looking infrared and hyperspectral synthetic aperture radar imagery. Current ATR algorithms are based on statistical, machine learning, information theoretic, theoretical, and neural networks.

We plan to develop advanced ATR algorithms that consist of fusion of multisensors and multi-algorithms. Several different ATR classifiers are being integrated in our laboratory. More research is needed to obtain optimal fusion methods for these ATR algorithms. Such hybrid algorithms could involve a single sensor or several sensors.

Related research areas include knowledge-based image-processing algorithms, model-based object recognition, statistical pattern recognition, Bayesian networks, machine learning, graph theory, fuzzy neural networks, and sensor fusion.